Two-step data clustering for improved intrusion detection system using CICIoT2023 dataset
Hadeel Qasem Gheni, Wathiq Laftah Al-Yaseen
Abstract
The issue of network security is an important and delicate issue when it comes to the privacy of organizations and individuals, especially when important and sensitive information is transmitted across these networks. The importance of intrusion detection systems, which is a very important component of protecting information and reducing the damage resulting from attacks and penetrations of network systems, has increased due to the adoption of the most recent regulations on advanced web services, whether government services, banking services, e-mail, or e-marketing. The goal of this paper is to construct an intrusion detection system using deep learning algorithms based on a new dataset named the CICIoT2023. The proposed intrusion detection model addresses challenges associated with intrusion detection datasets in terms of high dimensionality by adopting new methods to reduce their size and improve efficiency. A new clustering technique for intrusion detection datasets based on a new method combination between an optimization algorithm and static tools was proposed. The proposed model was evaluated to determine its efficiency using several evaluation measures. The results show that in comparison to earlier research conducted on the same datasets, the suggested model performs better in attack detection. As a result, the proposed model offers a high level of network security trust.